local-voice/tests/run_acceptance.py

#!/usr/bin/env python3
"""Create and optionally render fresh Local Voice acceptance plans."""

from __future__ import annotations

import argparse
import json
import subprocess
import sys
from pathlib import Path


ROOT = Path(__file__).resolve().parent.parent
RENDERER = ROOT / "scripts" / "local_voice.py"
CASES = (
    "donna",
    "chris-engineer",
    "grandpa-bomber",
    "val-holiday",
    "ryan-pitch-meeting",
)


def run(command: list[str]) -> None:
    subprocess.run(command, check=True)


def main() -> int:
    parser = argparse.ArgumentParser()
    parser.add_argument("--runtime-root", required=True, type=Path)
    parser.add_argument("--output-dir", required=True, type=Path)
    parser.add_argument("--device", choices=("metal", "cpu", "cuda"), default="metal")
    parser.add_argument("--render", action="store_true")
    parser.add_argument(
        "--voices",
        help="Comma-separated subset of profile IDs; defaults to every case.",
    )
    args = parser.parse_args()
    args.output_dir.mkdir(parents=True, exist_ok=True)
    selected_cases = (
        tuple(item.strip() for item in args.voices.split(",") if item.strip())
        if args.voices
        else CASES
    )
    unknown = sorted(set(selected_cases) - set(CASES))
    if unknown:
        parser.error("Unknown acceptance voice(s): " + ", ".join(unknown))
    results = []
    for voice in selected_cases:
        script = ROOT / "tests" / "acceptance" / f"{voice}.md"
        plan = args.output_dir / f"{voice}.json"
        audio = args.output_dir / f"{voice}.mp3"
        job_id = f"acceptance-v9-{voice}"
        run(
            [
                sys.executable,
                str(RENDERER),
                "plan",
                "--voice",
                voice,
                "--script",
                str(script),
                "--output-plan",
                str(plan),
                "--output-audio",
                str(audio),
                "--job-id",
                job_id,
            ]
        )
        if voice == "ryan-pitch-meeting":
            data = json.loads(plan.read_text(encoding="utf-8"))
            data["segments"] = [
                {
                    "id": "01-producer",
                    "role": "producer",
                    "text": "So, you have a Monday briefing for me?",
                    "context_after": "Yes sir, I do!",
                    "pause_after_ms": 95,
                },
                {
                    "id": "02-yes-sir",
                    "role": "writer",
                    "text": "Yes sir, I do!",
                    "fixed_asset": (
                        "voices/ryan-pitch-meeting/signature-exchanges/"
                        "yes-sir-i-do/lotr_two_towers.wav"
                    ),
                    "pause_after_ms": 55,
                },
                {
                    "id": "03-writer",
                    "role": "writer",
                    "text": (
                        "The Phoenix hardware shipped, but the final quantity "
                        "approval is still missing!"
                    ),
                    "context_before": "Yes sir, I do!",
                    "context_after": (
                        "Oh, paperwork blocking shipped hardware is tight!"
                    ),
                    "pause_after_ms": 95,
                },
                {
                    "id": "04-tight",
                    "role": "producer",
                    "text": "Oh, paperwork blocking shipped hardware is tight!",
                    "tts_text": "Oh, paperwork blocking shipped hardware",
                    "context_before": (
                        "The final quantity approval is still missing!"
                    ),
                    "context_after": "It sure is!",
                    "suffix_asset": (
                        "voices/ryan-pitch-meeting/signature-exchanges/"
                        "tight-endings/is/disclosure.wav"
                    ),
                    "suffix_gap_ms": 130,
                    "pause_after_ms": 95,
                },
                {
                    "id": "05-producer",
                    "role": "producer",
                    "text": (
                        "Is it going to be hard to clear that before noon?"
                    ),
                    "context_before": (
                        "Oh, paperwork blocking shipped hardware is tight!"
                    ),
                    "context_after": (
                        "Actually, it's gonna be super easy, barely an "
                        "inconvenience! Oh, really?"
                    ),
                    "pause_after_ms": 95,
                },
                {
                    "id": "06-super-easy",
                    "role": "writer",
                    "text": (
                        "Actually, it's gonna be super easy, barely an "
                        "inconvenience! Oh, really?"
                    ),
                    "fixed_asset": (
                        "voices/ryan-pitch-meeting/signature-exchanges/"
                        "super_easy_oh_really.wav"
                    ),
                    "pause_after_ms": 95,
                },
            ]
            plan.write_text(json.dumps(data, indent=2) + "\n", encoding="utf-8")
        run([sys.executable, str(RENDERER), "validate", str(plan)])
        if args.render:
            run(
                [
                    sys.executable,
                    str(RENDERER),
                    "render",
                    str(plan),
                    "--runtime-root",
                    str(args.runtime_root),
                    "--device",
                    args.device,
                    "--resume",
                ]
            )
        results.append({"voice": voice, "plan": str(plan), "audio": str(audio)})
    summary = args.output_dir / "acceptance-summary.json"
    existing = []
    if summary.exists():
        existing = json.loads(summary.read_text(encoding="utf-8"))
    merged = {item["voice"]: item for item in existing}
    merged.update({item["voice"]: item for item in results})
    summary.write_text(
        json.dumps([merged[key] for key in sorted(merged)], indent=2) + "\n",
        encoding="utf-8",
    )
    print(f"Acceptance summary: {summary}")
    return 0


if __name__ == "__main__":
    raise SystemExit(main())